Machine Learning Engineer Nanodegree

Convolutional Neural Networks

Project: Write an Algorithm for a Dog Identification App


Why We're Here

In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project.

Sample Dog Output

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!

The Road Ahead

We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.

  • Step 0: Import Datasets
  • Step 1: Detect Humans
  • Step 2: Detect Dogs
  • Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
  • Step 4: Use a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 5: Create a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 6: Write your Algorithm
  • Step 7: Test Your Algorithm

Step 0: Import Datasets

Import Dog Dataset

In the code cell below, we import a dataset of dog images. We populate a few variables through the use of the load_files function from the scikit-learn library:

  • train_files, valid_files, test_files - numpy arrays containing file paths to images
  • train_targets, valid_targets, test_targets - numpy arrays containing onehot-encoded classification labels
  • dog_names - list of string-valued dog breed names for translating labels
In [1]:
from sklearn.datasets import load_files       
from keras.utils import np_utils
import numpy as np
from glob import glob

# define function to load train, test, and validation datasets
def load_dataset(path):
    data = load_files(path)
    dog_files = np.array(data['filenames'])
    dog_targets = np_utils.to_categorical(np.array(data['target']), 133)
    return dog_files, dog_targets

# load train, test, and validation datasets
train_files, train_targets = load_dataset('dogImages/train')
valid_files, valid_targets = load_dataset('dogImages/valid')
test_files, test_targets = load_dataset('dogImages/test')

# load list of dog names
dog_names = [item[20:-1] for item in sorted(glob("dogImages/train/*/"))]

# print statistics about the dataset
print('There are %d total dog categories.' % len(dog_names))
print('There are %s total dog images.\n' % len(np.hstack([train_files, valid_files, test_files])))
print('There are %d training dog images.' % len(train_files))
print('There are %d validation dog images.' % len(valid_files))
print('There are %d test dog images.'% len(test_files))
Using TensorFlow backend.
There are 133 total dog categories.
There are 8351 total dog images.

There are 6680 training dog images.
There are 835 validation dog images.
There are 836 test dog images.

Import Human Dataset

In the code cell below, we import a dataset of human images, where the file paths are stored in the numpy array human_files.

In [2]:
import random
random.seed(8675309)

# load filenames in shuffled human dataset
human_files = np.array(glob("lfw/*/*"))
random.shuffle(human_files)

# print statistics about the dataset
print('There are %d total human images.' % len(human_files))
There are 13233 total human images.

Step 1: Detect Humans

We use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images. OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory.

In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.

In [3]:
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline                               

# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# load color (BGR) image
img = cv2.imread(human_files[3])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# find faces in image
faces = face_cascade.detectMultiScale(gray)

# print number of faces detected in the image
print('Number of faces detected:', len(faces))

# get bounding box for each detected face
for (x,y,w,h) in faces:
    # add bounding box to color image
    cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Number of faces detected: 3

Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.

In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.

Write a Human Face Detector

We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.

In [4]:
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
    img = cv2.imread(img_path)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray)
    return len(faces) > 0

(IMPLEMENTATION) Assess the Human Face Detector

Question 1: Use the code cell below to test the performance of the face_detector function.

  • What percentage of the first 100 images in human_files have a detected human face?
  • What percentage of the first 100 images in dog_files have a detected human face?

Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.

Answer:

The algorithm detected a human face in 99% of the images in "human_files".

The algorithm detected a human face in 11% of the images in "dog_files".

In [5]:
human_files_short = human_files[:100]
dog_files_short = train_files[:100]


## Test the performance of the face_detector algorithm 
## on the images in human_files_short and dog_files_short.

detected_face_human = np.array([1 if face_detector(i) else 0 for i in human_files_short])
detected_face_dog = np.array([1 if face_detector(i) else 0 for i in dog_files_short])

print('The algorithm detected a human face in {}% of the images in "human_files"'.format(np.sum(detected_face_human)))
print('The algorithm detected a human face in {}% of the images in "dog_files"'.format(np.sum(detected_face_dog)))
The algorithm detected a human face in 99% of the images in "human_files"
The algorithm detected a human face in 11% of the images in "dog_files"

Question 2: This algorithmic choice necessitates that we communicate to the user that we accept human images only when they provide a clear view of a face (otherwise, we risk having unneccessarily frustrated users!). In your opinion, is this a reasonable expectation to pose on the user? If not, can you think of a way to detect humans in images that does not necessitate an image with a clearly presented face?

Answer:

No, the condition that the user should provide only those images that have a clear view of face is not reasonable. There might be a number of scenarios where the image provided by the user does not meet this criteria (for example - the image contains the side view of the face or the image contains some noise), but, it's still a human face. Therefore, our algorithm should be able to detect whether there is a human face in the image or not, irrespective of the 'clarity' of the face.

We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on each of the datasets.

In [ ]:
## (Optional) Report the performance of another  
## face detection algorithm on the LFW dataset
### Feel free to use as many code cells as needed.

Step 2: Detect Dogs

In this section, we use a pre-trained ResNet-50 model to detect dogs in images. Our first line of code downloads the ResNet-50 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories. Given an image, this pre-trained ResNet-50 model returns a prediction (derived from the available categories in ImageNet) for the object that is contained in the image.

In [6]:
from keras.applications.resnet50 import ResNet50

# define ResNet50 model
ResNet50_model = ResNet50(weights='imagenet')

Pre-process the Data

When using TensorFlow as backend, Keras CNNs require a 4D array (which we'll also refer to as a 4D tensor) as input, with shape

$$ (\text{nb_samples}, \text{rows}, \text{columns}, \text{channels}), $$

where nb_samples corresponds to the total number of images (or samples), and rows, columns, and channels correspond to the number of rows, columns, and channels for each image, respectively.

The path_to_tensor function below takes a string-valued file path to a color image as input and returns a 4D tensor suitable for supplying to a Keras CNN. The function first loads the image and resizes it to a square image that is $224 \times 224$ pixels. Next, the image is converted to an array, which is then resized to a 4D tensor. In this case, since we are working with color images, each image has three channels. Likewise, since we are processing a single image (or sample), the returned tensor will always have shape

$$ (1, 224, 224, 3). $$

The paths_to_tensor function takes a numpy array of string-valued image paths as input and returns a 4D tensor with shape

$$ (\text{nb_samples}, 224, 224, 3). $$

Here, nb_samples is the number of samples, or number of images, in the supplied array of image paths. It is best to think of nb_samples as the number of 3D tensors (where each 3D tensor corresponds to a different image) in your dataset!

In [7]:
from keras.preprocessing import image                  
from tqdm import tqdm

def path_to_tensor(img_path):
    # loads RGB image as PIL.Image.Image type
    img = image.load_img(img_path, target_size=(224, 224))
    # convert PIL.Image.Image type to 3D tensor with shape (224, 224, 3)
    x = image.img_to_array(img)
    # convert 3D tensor to 4D tensor with shape (1, 224, 224, 3) and return 4D tensor
    return np.expand_dims(x, axis=0)

def paths_to_tensor(img_paths):
    list_of_tensors = [path_to_tensor(img_path) for img_path in tqdm(img_paths)]
    return np.vstack(list_of_tensors)

Making Predictions with ResNet-50

Getting the 4D tensor ready for ResNet-50, and for any other pre-trained model in Keras, requires some additional processing. First, the RGB image is converted to BGR by reordering the channels. All pre-trained models have the additional normalization step that the mean pixel (expressed in RGB as $[103.939, 116.779, 123.68]$ and calculated from all pixels in all images in ImageNet) must be subtracted from every pixel in each image. This is implemented in the imported function preprocess_input. If you're curious, you can check the code for preprocess_input here.

Now that we have a way to format our image for supplying to ResNet-50, we are now ready to use the model to extract the predictions. This is accomplished with the predict method, which returns an array whose $i$-th entry is the model's predicted probability that the image belongs to the $i$-th ImageNet category. This is implemented in the ResNet50_predict_labels function below.

By taking the argmax of the predicted probability vector, we obtain an integer corresponding to the model's predicted object class, which we can identify with an object category through the use of this dictionary.

In [8]:
from keras.applications.resnet50 import preprocess_input, decode_predictions

def ResNet50_predict_labels(img_path):
    # returns prediction vector for image located at img_path
    img = preprocess_input(path_to_tensor(img_path))
    return np.argmax(ResNet50_model.predict(img))

Write a Dog Detector

While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained ResNet-50 model, we need only check if the ResNet50_predict_labels function above returns a value between 151 and 268 (inclusive).

We use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).

In [9]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
    prediction = ResNet50_predict_labels(img_path)
    return ((prediction <= 268) & (prediction >= 151)) 

(IMPLEMENTATION) Assess the Dog Detector

Question 3: Use the code cell below to test the performance of your dog_detector function.

  • What percentage of the images in human_files_short have a detected dog?
  • What percentage of the images in dog_files_short have a detected dog?

Answer:

The algorithm detected a dog in 0% (none) of the images in "human_files".

The algorithm detected a human face in 100% of the images in "dog_files".

In [10]:
### Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.

detected_dog_human_files = np.array([1 if dog_detector(i) else 0 for i in human_files_short])
detected_dog_dog_files = np.array([1 if dog_detector(i) else 0 for i in dog_files_short])

print('The algorithm detected a dog in {}% of the images in "human_files"'.format(np.sum(detected_dog_human_files)))
print('The algorithm detected a human face in {}% of the images in "dog_files"'.format(np.sum(detected_dog_dog_files)))
The algorithm detected a dog in 0% of the images in "human_files"
The algorithm detected a human face in 100% of the images in "dog_files"

Step 3: Create a CNN to Classify Dog Breeds (from Scratch)

Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 1%. In Step 5 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.

Be careful with adding too many trainable layers! More parameters means longer training, which means you are more likely to need a GPU to accelerate the training process. Thankfully, Keras provides a handy estimate of the time that each epoch is likely to take; you can extrapolate this estimate to figure out how long it will take for your algorithm to train.

We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have great difficulty in distinguishing between a Brittany and a Welsh Springer Spaniel.

Brittany Welsh Springer Spaniel

It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).

Curly-Coated Retriever American Water Spaniel

Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.

Yellow Labrador Chocolate Labrador Black Labrador

We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.

Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!

Pre-process the Data

We rescale the images by dividing every pixel in every image by 255.

In [11]:
from PIL import ImageFile                            
ImageFile.LOAD_TRUNCATED_IMAGES = True                 

# pre-process the data for Keras
train_tensors = paths_to_tensor(train_files).astype('float32') / 255
valid_tensors = paths_to_tensor(valid_files).astype('float32') / 255
test_tensors = paths_to_tensor(test_files).astype('float32') / 255
100%|██████████| 6680/6680 [00:54<00:00, 123.43it/s]
100%|██████████| 835/835 [00:06<00:00, 137.79it/s]
100%|██████████| 836/836 [00:06<00:00, 167.20it/s]

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. At the end of your code cell block, summarize the layers of your model by executing the line:

    model.summary()

We have imported some Python modules to get you started, but feel free to import as many modules as you need. If you end up getting stuck, here's a hint that specifies a model that trains relatively fast on CPU and attains >1% test accuracy in 5 epochs:

Sample CNN

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. If you chose to use the hinted architecture above, describe why you think that CNN architecture should work well for the image classification task.

Answer:

The architecture mentioned above is a combination of Convolutional and MaxPooling layers, stacked alternately. The purpose of this kind of structure is to discover the spatial patterns contained in the image. The Convolutional layers are used to make the array (of image) deeper. The MaxPooling layers are used to decrease the spatial dimensions (width and height) of the image. Therefore, the spatial data is converted to a representation that encodes the content of the image (and all of the spatial information is eventually lost). Therefore, this architecture should work well for image classification.

In [12]:
from keras.layers import Conv2D, MaxPooling2D, GlobalAveragePooling2D
from keras.layers import Dropout, Flatten, Dense
from keras.models import Sequential

model = Sequential()

### Define the architecture.
model.add(Conv2D(filters=16, kernel_size=2, strides=1, padding='valid', activation='relu', input_shape=train_tensors.shape[1:]))
model.add(MaxPooling2D(pool_size=2, strides=2, padding='valid'))
model.add(Conv2D(filters=32, kernel_size=2, strides=1, padding='valid', activation='relu'))
model.add(MaxPooling2D(pool_size=2, strides=2, padding='valid'))
model.add(Conv2D(filters=64, kernel_size=2, strides=1, padding='valid', activation='relu'))
model.add(MaxPooling2D(pool_size=2, strides=2, padding='valid'))
model.add(GlobalAveragePooling2D())
model.add(Dense(133, activation='softmax'))

model.summary()
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_1 (Conv2D)            (None, 223, 223, 16)      208       
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 111, 111, 16)      0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 110, 110, 32)      2080      
_________________________________________________________________
max_pooling2d_3 (MaxPooling2 (None, 55, 55, 32)        0         
_________________________________________________________________
conv2d_3 (Conv2D)            (None, 54, 54, 64)        8256      
_________________________________________________________________
max_pooling2d_4 (MaxPooling2 (None, 27, 27, 64)        0         
_________________________________________________________________
global_average_pooling2d_1 ( (None, 64)                0         
_________________________________________________________________
dense_1 (Dense)              (None, 133)               8645      
=================================================================
Total params: 19,189.0
Trainable params: 19,189.0
Non-trainable params: 0.0
_________________________________________________________________

Compile the Model

In [13]:
model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])

(IMPLEMENTATION) Train the Model

Train your model in the code cell below. Use model checkpointing to save the model that attains the best validation loss.

You are welcome to augment the training data, but this is not a requirement.

In [14]:
from keras.callbacks import ModelCheckpoint  

epochs = 10

### Do NOT modify the code below this line.

checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.from_scratch.hdf5', 
                               verbose=1, save_best_only=True)

model.fit(train_tensors, train_targets, 
          validation_data=(valid_tensors, valid_targets),
          epochs=epochs, batch_size=20, callbacks=[checkpointer], verbose=1)
Train on 6680 samples, validate on 835 samples
Epoch 1/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.8845 - acc: 0.0107Epoch 00000: val_loss improved from inf to 4.86970, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 20s - loss: 4.8845 - acc: 0.0106 - val_loss: 4.8697 - val_acc: 0.0108
Epoch 2/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.8651 - acc: 0.0116Epoch 00001: val_loss improved from 4.86970 to 4.85774, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.8648 - acc: 0.0115 - val_loss: 4.8577 - val_acc: 0.0156
Epoch 3/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.8253 - acc: 0.0140Epoch 00002: val_loss improved from 4.85774 to 4.81503, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.8256 - acc: 0.0139 - val_loss: 4.8150 - val_acc: 0.0263
Epoch 4/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.7764 - acc: 0.0203Epoch 00003: val_loss improved from 4.81503 to 4.78817, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.7768 - acc: 0.0202 - val_loss: 4.7882 - val_acc: 0.0216
Epoch 5/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.7443 - acc: 0.0213Epoch 00004: val_loss improved from 4.78817 to 4.76822, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.7449 - acc: 0.0213 - val_loss: 4.7682 - val_acc: 0.0204
Epoch 6/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.7211 - acc: 0.0236Epoch 00005: val_loss improved from 4.76822 to 4.75381, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.7207 - acc: 0.0235 - val_loss: 4.7538 - val_acc: 0.0180
Epoch 7/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.6989 - acc: 0.0288Epoch 00006: val_loss improved from 4.75381 to 4.75291, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.6984 - acc: 0.0289 - val_loss: 4.7529 - val_acc: 0.0228
Epoch 8/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.6797 - acc: 0.0314Epoch 00007: val_loss improved from 4.75291 to 4.73619, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.6793 - acc: 0.0313 - val_loss: 4.7362 - val_acc: 0.0240
Epoch 9/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.6582 - acc: 0.0329Epoch 00008: val_loss improved from 4.73619 to 4.70836, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.6587 - acc: 0.0329 - val_loss: 4.7084 - val_acc: 0.0311
Epoch 10/10
6660/6680 [============================>.] - ETA: 0s - loss: 4.6375 - acc: 0.0345Epoch 00009: val_loss improved from 4.70836 to 4.70593, saving model to saved_models/weights.best.from_scratch.hdf5
6680/6680 [==============================] - 19s - loss: 4.6375 - acc: 0.0346 - val_loss: 4.7059 - val_acc: 0.0263
Out[14]:
<keras.callbacks.History at 0x7f13c10d2d68>

Load the Model with the Best Validation Loss

In [15]:
model.load_weights('saved_models/weights.best.from_scratch.hdf5')

Test the Model

Try out your model on the test dataset of dog images. Ensure that your test accuracy is greater than 1%.

In [16]:
# get index of predicted dog breed for each image in test set
dog_breed_predictions = [np.argmax(model.predict(np.expand_dims(tensor, axis=0))) for tensor in test_tensors]

# report test accuracy
test_accuracy = 100*np.sum(np.array(dog_breed_predictions)==np.argmax(test_targets, axis=1))/len(dog_breed_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
Test accuracy: 3.3493%

Step 4: Use a CNN to Classify Dog Breeds

To reduce training time without sacrificing accuracy, we show you how to train a CNN using transfer learning. In the following step, you will get a chance to use transfer learning to train your own CNN.

Obtain Bottleneck Features

In [17]:
bottleneck_features = np.load('bottleneck_features/DogVGG16Data.npz')
train_VGG16 = bottleneck_features['train']
valid_VGG16 = bottleneck_features['valid']
test_VGG16 = bottleneck_features['test']

Model Architecture

The model uses the the pre-trained VGG-16 model as a fixed feature extractor, where the last convolutional output of VGG-16 is fed as input to our model. We only add a global average pooling layer and a fully connected layer, where the latter contains one node for each dog category and is equipped with a softmax.

In [18]:
VGG16_model = Sequential()
VGG16_model.add(GlobalAveragePooling2D(input_shape=train_VGG16.shape[1:]))
VGG16_model.add(Dense(133, activation='softmax'))

VGG16_model.summary()
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
global_average_pooling2d_2 ( (None, 512)               0         
_________________________________________________________________
dense_2 (Dense)              (None, 133)               68229     
=================================================================
Total params: 68,229.0
Trainable params: 68,229.0
Non-trainable params: 0.0
_________________________________________________________________

Compile the Model

In [19]:
VGG16_model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])

Train the Model

In [20]:
checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.VGG16.hdf5', 
                               verbose=1, save_best_only=True)

VGG16_model.fit(train_VGG16, train_targets, 
          validation_data=(valid_VGG16, valid_targets),
          epochs=20, batch_size=20, callbacks=[checkpointer], verbose=1)
Train on 6680 samples, validate on 835 samples
Epoch 1/20
6460/6680 [============================>.] - ETA: 0s - loss: 12.2453 - acc: 0.1142Epoch 00000: val_loss improved from inf to 10.64673, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 12.1900 - acc: 0.1174 - val_loss: 10.6467 - val_acc: 0.2120
Epoch 2/20
6460/6680 [============================>.] - ETA: 0s - loss: 9.7245 - acc: 0.2899Epoch 00001: val_loss improved from 10.64673 to 9.76245, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 9.7319 - acc: 0.2903 - val_loss: 9.7624 - val_acc: 0.2850
Epoch 3/20
6480/6680 [============================>.] - ETA: 0s - loss: 9.0794 - acc: 0.3707Epoch 00002: val_loss improved from 9.76245 to 9.25823, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 9.0835 - acc: 0.3708 - val_loss: 9.2582 - val_acc: 0.3329
Epoch 4/20
6500/6680 [============================>.] - ETA: 0s - loss: 8.7861 - acc: 0.4089Epoch 00003: val_loss improved from 9.25823 to 9.17429, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.7813 - acc: 0.4087 - val_loss: 9.1743 - val_acc: 0.3581
Epoch 5/20
6480/6680 [============================>.] - ETA: 0s - loss: 8.6191 - acc: 0.4319Epoch 00004: val_loss improved from 9.17429 to 9.08216, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.6577 - acc: 0.4296 - val_loss: 9.0822 - val_acc: 0.3677
Epoch 6/20
6500/6680 [============================>.] - ETA: 0s - loss: 8.6010 - acc: 0.4386Epoch 00005: val_loss improved from 9.08216 to 9.04748, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.5739 - acc: 0.4395 - val_loss: 9.0475 - val_acc: 0.3545
Epoch 7/20
6520/6680 [============================>.] - ETA: 0s - loss: 8.4985 - acc: 0.4503Epoch 00006: val_loss improved from 9.04748 to 8.98510, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.4973 - acc: 0.4500 - val_loss: 8.9851 - val_acc: 0.3737
Epoch 8/20
6480/6680 [============================>.] - ETA: 0s - loss: 8.4611 - acc: 0.4540Epoch 00007: val_loss improved from 8.98510 to 8.90036, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.4455 - acc: 0.4551 - val_loss: 8.9004 - val_acc: 0.3832
Epoch 9/20
6520/6680 [============================>.] - ETA: 0s - loss: 8.3738 - acc: 0.4667Epoch 00008: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 8.3768 - acc: 0.4663 - val_loss: 8.9552 - val_acc: 0.3701
Epoch 10/20
6480/6680 [============================>.] - ETA: 0s - loss: 8.2277 - acc: 0.4793Epoch 00009: val_loss improved from 8.90036 to 8.74777, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.2638 - acc: 0.4774 - val_loss: 8.7478 - val_acc: 0.3928
Epoch 11/20
6460/6680 [============================>.] - ETA: 0s - loss: 8.2141 - acc: 0.4780Epoch 00010: val_loss improved from 8.74777 to 8.71299, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 8.1967 - acc: 0.4787 - val_loss: 8.7130 - val_acc: 0.3772
Epoch 12/20
6500/6680 [============================>.] - ETA: 0s - loss: 7.9857 - acc: 0.4834Epoch 00011: val_loss improved from 8.71299 to 8.53518, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.9486 - acc: 0.4850 - val_loss: 8.5352 - val_acc: 0.3964
Epoch 13/20
6480/6680 [============================>.] - ETA: 0s - loss: 7.7719 - acc: 0.5008Epoch 00012: val_loss improved from 8.53518 to 8.46277, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.7736 - acc: 0.5010 - val_loss: 8.4628 - val_acc: 0.3964
Epoch 14/20
6520/6680 [============================>.] - ETA: 0s - loss: 7.7299 - acc: 0.5092Epoch 00013: val_loss improved from 8.46277 to 8.39982, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.7309 - acc: 0.5093 - val_loss: 8.3998 - val_acc: 0.4144
Epoch 15/20
6560/6680 [============================>.] - ETA: 0s - loss: 7.7089 - acc: 0.5130Epoch 00014: val_loss improved from 8.39982 to 8.33469, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.7093 - acc: 0.5129 - val_loss: 8.3347 - val_acc: 0.4024
Epoch 16/20
6480/6680 [============================>.] - ETA: 0s - loss: 7.6592 - acc: 0.5150Epoch 00015: val_loss improved from 8.33469 to 8.30928, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.6380 - acc: 0.5165 - val_loss: 8.3093 - val_acc: 0.4168
Epoch 17/20
6480/6680 [============================>.] - ETA: 0s - loss: 7.4944 - acc: 0.5253Epoch 00016: val_loss improved from 8.30928 to 8.15445, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.4922 - acc: 0.5257 - val_loss: 8.1544 - val_acc: 0.4275
Epoch 18/20
6540/6680 [============================>.] - ETA: 0s - loss: 7.3974 - acc: 0.5303Epoch 00017: val_loss improved from 8.15445 to 8.08471, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.3877 - acc: 0.5307 - val_loss: 8.0847 - val_acc: 0.4144
Epoch 19/20
6500/6680 [============================>.] - ETA: 0s - loss: 7.1061 - acc: 0.5388Epoch 00018: val_loss improved from 8.08471 to 7.78145, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 7.0921 - acc: 0.5394 - val_loss: 7.7815 - val_acc: 0.4407
Epoch 20/20
6480/6680 [============================>.] - ETA: 0s - loss: 6.9025 - acc: 0.5563Epoch 00019: val_loss improved from 7.78145 to 7.68039, saving model to saved_models/weights.best.VGG16.hdf5
6680/6680 [==============================] - 1s - loss: 6.9054 - acc: 0.5560 - val_loss: 7.6804 - val_acc: 0.4419
Out[20]:
<keras.callbacks.History at 0x7f1398f3de48>

Load the Model with the Best Validation Loss

In [21]:
VGG16_model.load_weights('saved_models/weights.best.VGG16.hdf5')

Test the Model

Now, we can use the CNN to test how well it identifies breed within our test dataset of dog images. We print the test accuracy below.

In [22]:
# get index of predicted dog breed for each image in test set
VGG16_predictions = [np.argmax(VGG16_model.predict(np.expand_dims(feature, axis=0))) for feature in test_VGG16]

# report test accuracy
test_accuracy = 100*np.sum(np.array(VGG16_predictions)==np.argmax(test_targets, axis=1))/len(VGG16_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
Test accuracy: 46.6507%

Predict Dog Breed with the Model

In [23]:
from extract_bottleneck_features import *

def VGG16_predict_breed(img_path):
    # extract bottleneck features
    bottleneck_feature = extract_VGG16(path_to_tensor(img_path))
    # obtain predicted vector
    predicted_vector = VGG16_model.predict(bottleneck_feature)
    # return dog breed that is predicted by the model
    return dog_names[np.argmax(predicted_vector)]

Step 5: Create a CNN to Classify Dog Breeds (using Transfer Learning)

You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.

In Step 4, we used transfer learning to create a CNN using VGG-16 bottleneck features. In this section, you must use the bottleneck features from a different pre-trained model. To make things easier for you, we have pre-computed the features for all of the networks that are currently available in Keras:

The files are encoded as such:

Dog{network}Data.npz

where {network}, in the above filename, can be one of VGG19, Resnet50, InceptionV3, or Xception. Pick one of the above architectures, download the corresponding bottleneck features, and store the downloaded file in the bottleneck_features/ folder in the repository.

(IMPLEMENTATION) Obtain Bottleneck Features

In the code block below, extract the bottleneck features corresponding to the train, test, and validation sets by running the following:

bottleneck_features = np.load('bottleneck_features/Dog{network}Data.npz')
train_{network} = bottleneck_features['train']
valid_{network} = bottleneck_features['valid']
test_{network} = bottleneck_features['test']
In [11]:
### Obtain bottleneck features from another pre-trained CNN.

bottleneck_features = np.load('bottleneck_features/DogInceptionV3Data.npz')
train_InceptionV3 = bottleneck_features['train']
valid_InceptionV3 = bottleneck_features['valid']
test_InceptionV3 = bottleneck_features['test']

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. At the end of your code cell block, summarize the layers of your model by executing the line:

    <your model's name>.summary()

Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.

Answer:

I chose the InceptionV3 architecture for transfer learning. It increases the depth of the image to 2048, meaning that all of the content of the image is encoded in this representation.

Next is a GlobalAveragePooling layer to flatten the output of the InceptionV3 model. We can now add some fully connected layers in order to predict the breed of the dog in the input image.

I also added a hidden layer with 64 neurons with the ReLU activation function. This layer also has a dropout in which 50% of the neurons will be randomly shooted (dropped). This will prevent overfitting of the model.

Finally, there is the output layer with 133 neurons (1 for each class) with softmax activation, since we have multiple classes and softmax is what we use for multi-class classification.

The architecture is not very complicated (it only has around 140,000 training parameters), therefore the training time would not be very long, and the dropout layer will ensure that the model does not overfits.

In [12]:
from keras.models import Sequential
from keras.layers import GlobalAveragePooling2D
from keras.layers.core import Dense, Dropout
### Define your architecture.

Inception_model = Sequential()
Inception_model.add(GlobalAveragePooling2D(input_shape=train_InceptionV3.shape[1:]))
Inception_model.add(Dense(64, activation='relu'))
Inception_model.add(Dropout(0.5))
Inception_model.add(Dense(133, activation='softmax'))

Inception_model.summary()
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
global_average_pooling2d_1 ( (None, 2048)              0         
_________________________________________________________________
dense_1 (Dense)              (None, 64)                131136    
_________________________________________________________________
dropout_1 (Dropout)          (None, 64)                0         
_________________________________________________________________
dense_2 (Dense)              (None, 133)               8645      
=================================================================
Total params: 139,781.0
Trainable params: 139,781.0
Non-trainable params: 0.0
_________________________________________________________________

(IMPLEMENTATION) Compile the Model

In [13]:
### Compile the model.

Inception_model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])

(IMPLEMENTATION) Train the Model

Train your model in the code cell below. Use model checkpointing to save the model that attains the best validation loss.

You are welcome to augment the training data, but this is not a requirement.

In [84]:
from keras.callbacks import ModelCheckpoint
### Train the model.

checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.InceptionV3.hdf5', 
                               verbose=1, save_best_only=True)

Inception_model.fit(train_InceptionV3, train_targets, 
          validation_data=(valid_InceptionV3, valid_targets),
          epochs=20, batch_size=20, callbacks=[checkpointer], verbose=1)
Train on 6680 samples, validate on 835 samples
Epoch 1/20
6620/6680 [============================>.] - ETA: 0s - loss: 3.5656 - acc: 0.2154Epoch 00000: val_loss improved from inf to 1.49963, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 2s - loss: 3.5537 - acc: 0.2174 - val_loss: 1.4996 - val_acc: 0.6790
Epoch 2/20
6620/6680 [============================>.] - ETA: 0s - loss: 2.0653 - acc: 0.4678Epoch 00001: val_loss improved from 1.49963 to 0.91544, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 2.0635 - acc: 0.4680 - val_loss: 0.9154 - val_acc: 0.7605
Epoch 3/20
6620/6680 [============================>.] - ETA: 0s - loss: 1.5978 - acc: 0.5580Epoch 00002: val_loss improved from 0.91544 to 0.72193, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 1.5995 - acc: 0.5578 - val_loss: 0.7219 - val_acc: 0.7880
Epoch 4/20
6600/6680 [============================>.] - ETA: 0s - loss: 1.3456 - acc: 0.6229Epoch 00003: val_loss improved from 0.72193 to 0.65090, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 1.3454 - acc: 0.6240 - val_loss: 0.6509 - val_acc: 0.8060
Epoch 5/20
6620/6680 [============================>.] - ETA: 0s - loss: 1.2617 - acc: 0.6323Epoch 00004: val_loss improved from 0.65090 to 0.61959, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 1.2590 - acc: 0.6329 - val_loss: 0.6196 - val_acc: 0.8084
Epoch 6/20
6620/6680 [============================>.] - ETA: 0s - loss: 1.1842 - acc: 0.6560Epoch 00005: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 1.1821 - acc: 0.6567 - val_loss: 0.6236 - val_acc: 0.8024
Epoch 7/20
6600/6680 [============================>.] - ETA: 0s - loss: 1.1195 - acc: 0.6709Epoch 00006: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 1.1198 - acc: 0.6708 - val_loss: 0.6506 - val_acc: 0.8144
Epoch 8/20
6580/6680 [============================>.] - ETA: 0s - loss: 1.0676 - acc: 0.6884Epoch 00007: val_loss improved from 0.61959 to 0.61880, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 1.0688 - acc: 0.6876 - val_loss: 0.6188 - val_acc: 0.8180
Epoch 9/20
6600/6680 [============================>.] - ETA: 0s - loss: 1.0499 - acc: 0.6923Epoch 00008: val_loss improved from 0.61880 to 0.61044, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 1.0499 - acc: 0.6921 - val_loss: 0.6104 - val_acc: 0.8180
Epoch 10/20
6620/6680 [============================>.] - ETA: 0s - loss: 0.9878 - acc: 0.7045Epoch 00009: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9883 - acc: 0.7046 - val_loss: 0.6432 - val_acc: 0.8144
Epoch 11/20
6600/6680 [============================>.] - ETA: 0s - loss: 0.9761 - acc: 0.7068Epoch 00010: val_loss improved from 0.61044 to 0.59578, saving model to saved_models/weights.best.InceptionV3.hdf5
6680/6680 [==============================] - 1s - loss: 0.9761 - acc: 0.7057 - val_loss: 0.5958 - val_acc: 0.8311
Epoch 12/20
6620/6680 [============================>.] - ETA: 0s - loss: 0.9537 - acc: 0.7144Epoch 00011: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9540 - acc: 0.7147 - val_loss: 0.6357 - val_acc: 0.8144
Epoch 13/20
6620/6680 [============================>.] - ETA: 0s - loss: 0.9711 - acc: 0.7041Epoch 00012: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9702 - acc: 0.7039 - val_loss: 0.6261 - val_acc: 0.8311
Epoch 14/20
6620/6680 [============================>.] - ETA: 0s - loss: 0.8949 - acc: 0.7254Epoch 00013: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9024 - acc: 0.7243 - val_loss: 0.6468 - val_acc: 0.8263
Epoch 15/20
6600/6680 [============================>.] - ETA: 0s - loss: 0.9067 - acc: 0.7208Epoch 00014: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9097 - acc: 0.7205 - val_loss: 0.6504 - val_acc: 0.8156
Epoch 16/20
6540/6680 [============================>.] - ETA: 0s - loss: 0.9066 - acc: 0.7292Epoch 00015: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.9088 - acc: 0.7289 - val_loss: 0.6264 - val_acc: 0.8311
Epoch 17/20
6580/6680 [============================>.] - ETA: 0s - loss: 0.8919 - acc: 0.7237Epoch 00016: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.8966 - acc: 0.7232 - val_loss: 0.7086 - val_acc: 0.8180
Epoch 18/20
6600/6680 [============================>.] - ETA: 0s - loss: 0.8418 - acc: 0.7444Epoch 00017: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.8487 - acc: 0.7430 - val_loss: 0.6268 - val_acc: 0.8371
Epoch 19/20
6600/6680 [============================>.] - ETA: 0s - loss: 0.8232 - acc: 0.7488Epoch 00018: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.8248 - acc: 0.7490 - val_loss: 0.6704 - val_acc: 0.8287
Epoch 20/20
6620/6680 [============================>.] - ETA: 0s - loss: 0.8459 - acc: 0.7408Epoch 00019: val_loss did not improve
6680/6680 [==============================] - 1s - loss: 0.8454 - acc: 0.7412 - val_loss: 0.6558 - val_acc: 0.8323
Out[84]:
<keras.callbacks.History at 0x7f0c2c975d68>

(IMPLEMENTATION) Load the Model with the Best Validation Loss

In [14]:
### Load the model weights with the best validation loss.

Inception_model.load_weights('saved_models/weights.best.InceptionV3.hdf5')

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Ensure that your test accuracy is greater than 60%.

In [86]:
### Calculate classification accuracy on the test dataset.

# get index of predicted dog breed for each image in test set
Inception_predictions = [np.argmax(Inception_model.predict(np.expand_dims(feature, axis=0))) for feature in test_InceptionV3]

# report test accuracy
test_accuracy = 100*np.sum(np.array(Inception_predictions)==np.argmax(test_targets, axis=1))/len(Inception_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
Test accuracy: 79.6651%

(IMPLEMENTATION) Predict Dog Breed with the Model

Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan_hound, etc) that is predicted by your model.

Similar to the analogous function in Step 5, your function should have three steps:

  1. Extract the bottleneck features corresponding to the chosen CNN model.
  2. Supply the bottleneck features as input to the model to return the predicted vector. Note that the argmax of this prediction vector gives the index of the predicted dog breed.
  3. Use the dog_names array defined in Step 0 of this notebook to return the corresponding breed.

The functions to extract the bottleneck features can be found in extract_bottleneck_features.py, and they have been imported in an earlier code cell. To obtain the bottleneck features corresponding to your chosen CNN architecture, you need to use the function

extract_{network}

where {network}, in the above filename, should be one of VGG19, Resnet50, InceptionV3, or Xception.

In [15]:
### A function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.

from extract_bottleneck_features import *

def Inception_predict_breed(img_path):
    # extract bottleneck features
    bottleneck_feature = extract_InceptionV3(path_to_tensor(img_path))
    # obtain predicted vector
    predicted_vector = Inception_model.predict(bottleneck_feature)
    # return dog breed that is predicted by the model
    return dog_names[np.argmax(predicted_vector)]

Step 6: Write your Algorithm

Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,

  • if a dog is detected in the image, return the predicted breed.
  • if a human is detected in the image, return the resembling dog breed.
  • if neither is detected in the image, provide output that indicates an error.

You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and dog_detector functions developed above. You are required to use your CNN from Step 5 to predict dog breed.

Some sample output for our algorithm is provided below, but feel free to design your own user experience!

Sample Human Output

(IMPLEMENTATION) Write your Algorithm

In [16]:
###  Algorithm.
from IPython.display import display
from IPython.display import Image
def what_dog_is_this(img_path):
    if dog_detector(img_path):
        dog_breed = Inception_predict_breed(img_path)
        img = cv2.imread(img_path)
        plt.imshow(img)
        print('\nIt\'s a {}'.format(dog_breed))
        
    elif face_detector(img_path):
        print('Hey, you are a Human!')
        img = Image(img_path)
        display(img)
        
        resembled_breed = Inception_predict_breed(img_path)
        print('\nAnd if someone asks you what dog breed do you resemble.. \n\nYou\'re like {}'.format(resembled_breed))
        
        temp = [i for i in train_files if resembled_breed in i]
        if len(temp) > 0:
            original_dog = temp[0]
            print('\nAnd here is how the dog you look like, looks ---:')
            
            img2 = Image(original_dog)
            display(img2)
            
    else:
        print('Sorry! This is neither a dog nor a human -- an alien maybe... :P')
        img = cv2.imread(img_path)
        plt.imshow(img)

Step 7: Test Your Algorithm

In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?

(IMPLEMENTATION) Test Your Algorithm on Sample Images!

Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.

Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.

Answer:

The algorithm works perfectly. It gave the correct predictions for all the images of dogs that were provided as input. When I gave as input the image of a bottle, the algorithm was able to identify that the image provided had neither a dog nor a human face, and gave the appropriate error message. For all the human faces that I gave as input, the algorithm was able to identify the resembling breeds and surprisingly, I think that the predictions were somehow accurate :P. So, I think, the output was much better than I expected. But, there are a numerous amount of improvements that can be made to the algorithm. Some of them are:

  • A major improvement could be made in the Model used for prediction. By tuning the hyperparameters of the model, adding/reducing the number of layers or the number of neurons in each layer, or the activation function for each of the hidden layer, we can drastically improve the predictions made by our model.

  • If the image contains both a human face, and a dog, then the algorithm will detect the dog and predict its breed. It will not even check whether there is a human face in the image or not. The algorithm can be modified in such a way the if there are both a human face and a dog in the image, it would prompt the user whether it should consider the dog or the human face.

  • The algorithm could be made more user friendly. There can be an option for the user to use a web-camera to take a picture of him/her in real-time and use the algorithm to find the breed of the dog that they resemble the most.

  • Also, if a human face is given as input and the algorithm predicts its breed, it should also output how much that human resembles the predicted breed (eg - 48% of your face resembles a Beagle).

In [17]:
# Execution of the algorithm

what_dog_is_this(r'images/Labrador_retriever_06457.jpg')
It's a Labrador_retriever
In [130]:
what_dog_is_this(r'images/Labrador_retriever_06449.jpg')
It's a Labrador_retriever
In [131]:
what_dog_is_this(r'images/Curly-coated_retriever_03896.jpg')
It's a Curly-coated_retriever
In [132]:
what_dog_is_this(r'images/Brittany_02625.jpg')
It's a Brittany
In [133]:
what_dog_is_this(r'images/American_water_spaniel_00648.jpg')
It's a American_water_spaniel
In [134]:
what_dog_is_this(r'images/Welsh_springer_spaniel_08203.jpg')
It's a Welsh_springer_spaniel
In [135]:
what_dog_is_this(r'images/sample dog 1.jpg')
It's a Chow_chow
In [18]:
what_dog_is_this(r'images/sample dog 2.jpg')
It's a English_cocker_spaniel
In [20]:
what_dog_is_this(r'images/sample dog 3.jpg')
It's a Beagle
In [21]:
what_dog_is_this(r'images/sample dog 4.jpg')
It's a Tibetan_mastiff
In [149]:
what_dog_is_this(r'images/Mrinal Jain.jpg')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Pekingese

And here is how the dog you look like, looks ---:
In [150]:
what_dog_is_this(r'images/Kausher Karim.jpg')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Dandie_dinmont_terrier

And here is how the dog you look like, looks ---:
In [151]:
what_dog_is_this(r'images/Angshuman Mazumdar.PNG')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Greater_swiss_mountain_dog

And here is how the dog you look like, looks ---:
In [152]:
what_dog_is_this(r'images/Sanket Shikhar.jpg')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Akita

And here is how the dog you look like, looks ---:
In [153]:
what_dog_is_this(r'images/Katappa.jpg')
Hey, you are a Human!
IOPub data rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_data_rate_limit`.
And if someone asks you what dog breed do you resemble.. 

You're like Chinese_crested

And here is how the dog you look like, looks ---:
In [19]:
what_dog_is_this(r'images/Rishabh Rohan.jpg')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Collie

And here is how the dog you look like, looks ---:
In [154]:
what_dog_is_this(r'images/Danish.PNG')
Hey, you are a Human!
And if someone asks you what dog breed do you resemble.. 

You're like Portuguese_water_dog

And here is how the dog you look like, looks ---:
In [155]:
what_dog_is_this(r'images/bottle.jpg')
Sorry! This is neither a dog nor a human -- an alien maybe... :P